The Pentagon quietly funneled millions into a stealth Arlington startup called Twenty — a company that promises to automate hacking at scale and has reportedly landed a Cyber Command contract worth up to $12.6 million while picking up VC backing from In-Q-Tel and other big names. That revelation should alarm every American who believes in accountable government, because this is offensive cyber capability built in the shadows of our own capital. Taxpayer dollars deserve sunlight, not secrecy when we’re funding tools that can infiltrate networks around the world.
Job postings and the company’s own language make clear Twenty isn’t building defensive malware detectors — it’s designing AI agents to automate attack paths, craft convincing online personas, and run continuous operations across hundreds of targets. This isn’t theoretical; their recruiting ads explicitly ask for offensive cyber researchers and AI engineers to build agent-based frameworks and persona development tools. Outsourcing the act of breaking into systems to autonomous software raises obvious ethical, legal, and operational questions that our leaders have not adequately answered.
The worldwide stakes are already higher because private companies like Anthropic recently disclosed a September 2025 campaign in which a state-backed actor used agentic AI to perform the bulk of a cyber-espionage operation, reportedly executing 80–90 percent of tactical tasks and targeting about 30 organizations. If adversaries can cheaply scale intrusions with AI agents, the battlefield now includes our financial institutions, supply chains, and critical infrastructure — and Americans will pay the price in lost jobs, stolen intellectual property, and threats to national security. We must face that reality squarely instead of celebrating technical novelty.
Yes, America must maintain military superiority, but there’s a difference between preparedness and recklessness. The Department of Defense has poured large sums into frontier AI work with major labs, and the temptation to let private contractors quietly push the envelope is real — especially when those contractors are staffed by former military and intelligence officials comfortable operating beyond public view. The same entrepreneurs who served our country deserve oversight when their work creates new risks to civil liberties, international law, and the safety of ordinary Americans.
Conservatives should demand a clear policy: Congress must hold hearings, require declassification where feasible, and legislate strict guardrails on offensive AI cyber operations so these tools cannot be repurposed against American citizens or misused by partisans in future administrations. We can’t allow mission creep to convert tools meant for external threats into ministries of domestic surveillance or political warfare. Real security is built on transparency, rule of law, and a public that trusts its leaders to use force responsibly.
At the same time, we must push for parity: invest more in hardened defenses for our own networks, incentivize private-sector cyber hygiene, and make sure our intelligence community isn’t outpaced by adversaries weaponizing AI. Throwing money at stealth startups without oversight is a gamble the American people shouldn’t have to play. If the administration wants extraordinary authority to field autonomous cyber weapons, they should convince the American people and their elected representatives first — not hide behind classified contracts and venture capital cheerleaders.
Patriots who love freedom and enterprise can support a strong, smart posture that outcompetes China and other threats without sacrificing the liberties that make this country great. That means insisting on accountability, demanding congressional oversight, and refusing to let technological progress become a cover for secrecy and abuse. Hardworking Americans built this country; we won’t let bureaucrats hand it over to an unaccountable AI arms race.
